Opinion

Everything is random: Why history’s overrated for risk management

When choosing to invest in a security, hedge fund or another asset type, most investors generally look at the historical record of performance. For some types of investments or assets, historical performance is a mandatory consideration. Banks’ market risk analysis generally looks at the immediate history of the market as a gauge for how much value is likely to be lost in any given day.

But what seems like a rational financial theory — that history is a guide to the future — is in direct contrast to this modern psychological theory: What is actually random masquerades in our perceptions as the inevitable. So does history provide any guide to future performance? Maybe not.

The question of randomness versus order was unpacked in a recent podcast featuring the psychologist Nicholas Epley, author of the book “Mindwise: Why We Misunderstand What Others Think, Believe, Feel, and Want.” He used a sports example in which broadcasters typically display a firm belief that a “hot hand” or “winning streak” is evidence of a team or player being more likely to enjoy continued success during the streak, versus at other times. But is this really true?

Red dice

To look at the question more deeply, consider the idea that people mistakenly attach some kind of order to perceived randomness.

In another example, Epley discussed an experiment in which a magician asks people to toss a coin 20 times and write down the results. He asks another person to make up the results of 20 coin tosses and write down those fake results too. The magician’s challenge is to tell the fake from the real. Apparently, this is an easy trick. Why? The person faking the results imagines that in a real coin-toss situation, where there is a 50-50 probability of the results being heads or tails, the result will flip between the two on a more or less sequential basis. But that assumption is obviously false. In reality, random coin tosses will generally result in long runs for either heads or tails across the 20 coin tosses, just like a basketball player making 20 straight shots — or missing 10 — is more random than it seems.

How does this help us understand the future performance dilemma for financial traders?

It is understandable that in finance, those with a rock-solid performance history are attractive to a firm. There is a natural predilection for the hot hand. Someone who has done well in the past holds moneymaking promise for the future. The perception is that normally traders alternate between success and failure from one day to the next, so hot streaks are seen to be the result of extraordinary talent. A very long run of success appears to us mortals like a hot hand, providing evidence of success in the immediate future. This is demonstrated by the activity of the stock market. When the market has gone up for a long period of time, people start talking about it in everyday conversation and one feels foolish for not participating in it.

Bernie Madoff took advantage of this tendency among investors. His prospectus and marketing materials presented to his banks and clients the picture of an unbroken record of success. Similarly, leading up to the 2008 financial crisis, banks invested more and more deeply into the housing market — pulled in by the apparent surety of returns. When Goldman Sachs had begun to pull out and hedge its housing-related positions in 2006, many of its peers were still doubling down. Market observers have noted that there has been a similar rush to fund the student debt market over the past decade, leading to growth of over 170% over the past decade to a huge $1.4 trillion, which exceeds car loans or credit card debt. Some argue a similar trend may be at work in the funding of fintech startups.

Clearly, investors cannot afford to make investments based on past records of success. A more defensive, and perhaps effective, approach is to always assume the worst — that a winning streak will end tomorrow, or that the worst-scenarios are always more likely to occur than the best-case scenarios. This too can be based on history, as long as the risk model looks at a longer historical period.

Sometimes, too, history is a less reliable guide than simply imagining the unknown. Take the Y2K project, for instance. Farhad Manjoo, writing in The New York Times, noted that it was only putting the fear of God in people, about the potential effects of computers not being programmed to switch internal clocks to the year 2000, which drove people to do something about it. In the end, there was no Y2K crisis. “Y2K is one of the precious few examples where we mobilized to fight something on the horizon,” Manjoo wrote.

Similarly, if we want to make the wisest choices about current financial services strategy, we should look at the bad possible outcomes and not just the recent trajectory of history.

This article originally appeared in American Banker.
For reprint and licensing requests for this article, click here.
Risk management Predictive modeling Data quality
MORE FROM NATIONAL MORTGAGE NEWS